Be patient with this method, sometimes you have to experiment with the image strength to find a sweet spot, certain filters work better with certain images and experiment with the samplers.
@@theeirishstud Good to hear bud! when I first started messing around with this technique, I was getting horrendous results 😆 but it really is a balancing act. One of the advantages of running it locally on Automatic 1111 is that I can train with a service like Open AI's Photobooth, download the model (ckpt file) install it and I can make anything I want with that specific model. I really hope PAI looks at providing a similar service where we can train our own custom models. That would be next level!
@@MonzonMedia yeah they're really stoking the fires of creativity. I agree with that too. My PC can't do anything locally. They're doing good for how fast stuff rolls out. Balancing new features & website improvements.
That’s awesome! Yes nowadays you just need a camera phone and ai them you’re good to go! I have a video to work on, on this very topic. Appreciate the feedback!
@@williamflaherty3168 I'll definitely be doing a video on that very soon but for now you can start with this on the official playgroundai channel. th-cam.com/video/dkUjHPQ4nHA/w-d-xo.html
@MonzonMedia according to Perplexity AI: The name "Heun" can be pronounced as "h y OO n" in the US or "H OY NE" in Chicago, IL, USA, with the origin being American or German. In German, it is pronounced as "hoin".
What I am still looking for, is a way to use an image of for example a character and then change the posing, like turn it into a side profile or back profile, but keep hairstyle, clothes etc the same. This would be great for a visual novel. For dreambooth you need pictures of the same thing from various angles, but you can't quite do that with a picture you create in playground. I did manage to turn a head front profile a bit into a back profile with playground but either it keeps the rest of the body the same or it changes too much.
It's a bit tricky without being able to use Dreambooth or a service like Open AI's Photobooth, however I do have a work around hack that works pretty well for generating the same face through prompts only. I will post hopefully tomorrow.
Great vid ! Plz plz plz show how to do this on mid journey ..I keep getting deformed images . I’m trying to start a business where you can turn ppls photos of themselves or their pets into anime , superhero’s etc ….this has been the closest video I’ve found thus far on here
Hey there, appreciate it! OK, this is going to be a bit long winded....Currently you can't do this in Midjourney. Technically you can upload images of people to use as reference and you can get a "likeness" of the person but there is some bias to how the AI was trained. I've tried numerous images using real photos I've taken of people and it works ok but not exactly what you need. Since you want to do this as a business (I'm doing something similar), here's my advice to you. You should look into running stable diffusion locally with either Automatic1111 or Invoke AI. These are web UI's that run locally on your computer however you need a decent dedicated graphics card. Not sure what system you have. Option 1 is to use Dreambooth to train models locally, the downside of this is you need a beefy graphics card like an RTX 3090 with 24GB of RAM. Most people can't afford a card like that including myself. So what I've done instead (option 2) is use a service like Open AI's Photobooth. You can train models based on pictures you upload (usually 20-30 images) and from there you can generate images on the site. You also get credits when you buy the service which is very reasonable starting at $10. They also provide a CKPT file which you can download and install it locally on Automatic1111 or Invoke AI. From there you can generate unlimited images of that trained model. Personally when I do this I would then use an editor like Photoshop for example to do some final touch ups. If you are serious about making this a business, I highly suggest you consider running it locally (until the tech is more accessible). It will be much easier down the road and you will have more freedom to generate any type of image you want! Hope that helps!
@@MonzonMedia wow ! I can’t thank you enough for taking your time to give me this thorough response ! Now I know! And will be taking your valuable advice ! Thank you Thank you Thank you ! Very much appreciated! Keep up the great vids ! You have a loyal subscriber here 🙌🏼
@@adrewbateman1760 You are so welcome! I'm glad to help any way I can. I truly appreciate the support. Feel free to contact me if you have more questions. Between you and I **spoiler alert** I'm working on content on this very topic. 🤫😁
@@milestrombley1466 try this In the beginning of the video th-cam.com/video/zqUaZu4ODQw/w-d-xo.html or you could do a hack blending prompt and image th-cam.com/video/Bk-FhQsbknM/w-d-xo.html
No need for pro, you can do this with the free account. For some images you really have to experiment with what settings work. Sometimes if you change the sampler, it will work better or you may have to play with the image strength settings.
I have problems starting with a closeup portrait but want a full body picture as an end result. Playground AI sticks on the upper body. Can you give me a hint, please?
That's a bit tricky since there is no outpainting feature but there is a work around. You can download the image, recompose it in another editor like photoshop, pixlr or even Canva. Then re-upload and use inpaint to add the rest. I'll be doing a video on this soon. It's not a fool proof method but it can work to some extent.
Hey bud, any chances you are using the Playground V1 model or Stable Diffusion 2.1? All the filters are available with Stable Diffusion 1.5 at the moment. Only select ones for Playground V1 model. Let me know if that helps!
Hello again Julia! Yes absolutely and I am preparing a video for that as well. However the benefit of Leonardo AI is that you can train your own model based on real people if you have the photos to use for training. Check it out here th-cam.com/video/oNajDvx3_GY/w-d-xo.html and I also cover some updates on this follow up video on the second half of the video th-cam.com/video/ysg06oLGA94/w-d-xo.html
When I open my FILTER, the samples are different from yours. My question is: How can I upload other samples of images to change the styles in my filter?
I’m assuming you are using a reference image? You can start with a filter like analog style, in the negative prompt area you can put in cartoon, 2D, drawing etc. In your main prompt it can help to put photo of, portrait of, photorealistic. Might take a few tries. Hope that helps!
Great question! Technically, they do the same thing, isolate an area. I tend to say mask more because that's term we use for photo editing. But in the AI world we say inpaint. I'm sure others would have their own opinion on it but at the end of the day, it's the same thing. Isolating an area. Hope that makes sense?
Sometimes it takes the right image. I have yet to find an image that can do me well...just looks weird hahaha! You can keep trying with another image and works best with head and shoulder or half body shots. It's much better to be able to train a model with many images so that you can get the likeness of the person easier. Have you tried Leonardo AI? They have the ability to train. I did a video on it here. th-cam.com/video/oNajDvx3_GY/w-d-xo.html and at the end of this video I gave some updates. th-cam.com/video/ysg06oLGA94/w-d-xo.html. The free tier is a bit limited though.
Hello and thanks for a comprehensible video. Actually I am going crazy with this, as it is AI it should have a logical pattern of behaving. I did created a picture of a blueberry monster B for Berry and it was supposed to hold in the hand a banana. He was not, the bananas were on the ground and one above the head. I could erase them easily but then I spent five hours trying to add it to the picture so it would be in his hand. NO way. I was able to put bananas in different places, to place different objects in the picture. I tried image to image, inpainting and I am kinda deceived because I could do it easily in ten minutes in Photoshop and I lost all my day, in expectation that I would learn how to do adding and removing easily. Do you have some video on this topic, like detailed one, dealing with what can be established wrongly if it is so difficult to add a simple banana to the already perfect picture.
Thanks for checking out the video and the feedback. AI in general has trouble holding objects in hands. Usually inpainting can fix it or get close to what you want it to look like but I've found that it's sometimes quicker to do it in post. Also, many of the web based platforms lack certain features like ControlNet for example which you can run locally on a PC and this allows you make specific poses. I'd love to do a video on this topic actually.
There isn't much of a difference in how the product performs between the free and paid options, mostly it's the step count (quality) that is limited to 50 which is more than enough for most images.
Then instructions seemed pretty straight forward, but my first image was NSFW even thought it was just a regular pic of me in a gray t-shirt. Then the generate button stopped working. Tried to reload but its like it doesn't exist. Oh well.
Sure it is, I use it all the time. After you generate an image on the top corner is your actions menu, there you will see inpaint. If it's greyed out you have to turn off the filter.
aaahhhh I see, well in that case just use words that you don't want to see in the image. Whatever results are not to your liking just add it as you go, like blurry, warped (for things like tires). It's a bit difficult to come up with one list for objects as there are many different objects.
Be patient with this method, sometimes you have to experiment with the image strength to find a sweet spot, certain filters work better with certain images and experiment with the samplers.
Ok sir thank you
@@souravbijlwan5029 My pleasure!
Agreed. Took me a while but finaly I got a good retro anime style profile picture after a day or 2.
@@theeirishstud Good to hear bud! when I first started messing around with this technique, I was getting horrendous results 😆 but it really is a balancing act. One of the advantages of running it locally on Automatic 1111 is that I can train with a service like Open AI's Photobooth, download the model (ckpt file) install it and I can make anything I want with that specific model. I really hope PAI looks at providing a similar service where we can train our own custom models. That would be next level!
@@MonzonMedia yeah they're really stoking the fires of creativity. I agree with that too. My PC can't do anything locally. They're doing good for how fast stuff rolls out. Balancing new features & website improvements.
With your tutorials, I created a studio quality photorealistic portrait of myself in Giorgio Armani suit from a simple selfie. Thanks a lot!
That’s awesome! Yes nowadays you just need a camera phone and ai them you’re good to go! I have a video to work on, on this very topic. Appreciate the feedback!
share tutorial for photorealistic sir
Photorealistic images or making photorealistic avatars?
@@MonzonMedia A studio photorealistic portrait of upper body with dark greyish background and 3 point lighting, wearing perfect and realistic suit.
@@williamflaherty3168 I'll definitely be doing a video on that very soon but for now you can start with this on the official playgroundai channel. th-cam.com/video/dkUjHPQ4nHA/w-d-xo.html
I love this method. I was doing it with editing and not inpainting. This gives better results. Thanks
You're welcome! Yes much better to use inpainting before the edit function. 👍
I always learn something new watching your videos. Thanks for doing them.
My pleasure bud and I'm glad you are getting something out of them! Appreciate the support Richard! 🙌
bro thank you.. as a newbie in ai world and in playground specifically, this really helps me a lot ❤️
You're welcome Zaid, I'm glad I could help! More vids to come on playground. 🙌👍
Ahhh that’s how you do it! That’s going to unlock so much more for me. Thank you 🙏 Love your work
You’re welcome 😊 Thank you for the support!
The best video about image to image! 👏😎
I appreciate that! Thank you! I'm glad it was helpful. Are you only using Playground or do you use any others?
@MonzonMedia according to Perplexity AI:
The name "Heun" can be pronounced as "h y OO n" in the US or "H OY NE" in Chicago, IL, USA, with the origin being American or German. In German, it is pronounced as "hoin".
I appreciate it! I learned the correct way shortly after this video but good to get some verification! Where do they find these names? hahaha!
What I am still looking for, is a way to use an image of for example a character and then change the posing, like turn it into a side profile or back profile, but keep hairstyle, clothes etc the same. This would be great for a visual novel. For dreambooth you need pictures of the same thing from various angles, but you can't quite do that with a picture you create in playground. I did manage to turn a head front profile a bit into a back profile with playground but either it keeps the rest of the body the same or it changes too much.
It's a bit tricky without being able to use Dreambooth or a service like Open AI's Photobooth, however I do have a work around hack that works pretty well for generating the same face through prompts only. I will post hopefully tomorrow.
Funny, I was also doing some AI stuff using a different app, and it got me hooked, lol.
It's very addicting bud, especially with PAI since they offer a very generous amount of free images. (1K). What app where you using?
@Monzon Media . Ouuu. I'm using Prequel hehe. 😁
guess i have to watch this again
Just have to be patient with the process and find settings that work for you. Thanks for checking it out.
I'm also 😅
Need part 2
You are really a awesome guy
♥️ from India
Hello my friend from India! That's awesome! 🙌 Part 2 coming soon! Hang tight. 👍
My inpaint function is greyed out. Is this only for playground Pro? Thanks in advance.
Hi Eddie, most likely you will have to take off the filter before you can inpaint. Give that a shot. 👍
Excellent overview
Thank you so much! I appreciate it. 👍🏼
Most excellent, thank you!!
You're very welcome! 👍
Make more videos on Avatar in Playground AI
Sure but what else should I cover? More styles?
Great vid ! Plz plz plz show how to do this on mid journey ..I keep getting deformed images . I’m trying to start a business where you can turn ppls photos of themselves or their pets into anime , superhero’s etc ….this has been the closest video I’ve found thus far on here
Hey there, appreciate it! OK, this is going to be a bit long winded....Currently you can't do this in Midjourney. Technically you can upload images of people to use as reference and you can get a "likeness" of the person but there is some bias to how the AI was trained. I've tried numerous images using real photos I've taken of people and it works ok but not exactly what you need. Since you want to do this as a business (I'm doing something similar), here's my advice to you. You should look into running stable diffusion locally with either Automatic1111 or Invoke AI. These are web UI's that run locally on your computer however you need a decent dedicated graphics card. Not sure what system you have. Option 1 is to use Dreambooth to train models locally, the downside of this is you need a beefy graphics card like an RTX 3090 with 24GB of RAM. Most people can't afford a card like that including myself. So what I've done instead (option 2) is use a service like Open AI's Photobooth. You can train models based on pictures you upload (usually 20-30 images) and from there you can generate images on the site. You also get credits when you buy the service which is very reasonable starting at $10. They also provide a CKPT file which you can download and install it locally on Automatic1111 or Invoke AI. From there you can generate unlimited images of that trained model. Personally when I do this I would then use an editor like Photoshop for example to do some final touch ups. If you are serious about making this a business, I highly suggest you consider running it locally (until the tech is more accessible). It will be much easier down the road and you will have more freedom to generate any type of image you want! Hope that helps!
@@MonzonMedia wow ! I can’t thank you enough for taking your time to give me this thorough response ! Now I know! And will be taking your valuable advice ! Thank you Thank you Thank you ! Very much appreciated! Keep up the great vids ! You have a loyal subscriber here 🙌🏼
@@adrewbateman1760 You are so welcome! I'm glad to help any way I can. I truly appreciate the support. Feel free to contact me if you have more questions. Between you and I **spoiler alert** I'm working on content on this very topic. 🤫😁
Спасибо, с удовольствием посмотрела этот урок!
Подписалась, буду частым гостем на вашем канале.
I’m so glad to hear that and thank you! Really appreciate the support! 🙌🏼
Can you make a video on how to turn a character or photo into an anthropomorphic character in Playground AI, please?
Like a werewolf for an example? Because you can just add anthropomorphic in the prompt. Or were you thinking of something else? Let me know!
A werewolf or a furry cartoon character. Sometimes the AI doesn't understand what an anthropomorphic is. But I will give it a shot. Thanks.
@@milestrombley1466 try this In the beginning of the video th-cam.com/video/zqUaZu4ODQw/w-d-xo.html or you could do a hack blending prompt and image th-cam.com/video/Bk-FhQsbknM/w-d-xo.html
didnt got the result everything is mesed up. i think we need to buy the pro version cause many things are missing
No need for pro, you can do this with the free account. For some images you really have to experiment with what settings work. Sometimes if you change the sampler, it will work better or you may have to play with the image strength settings.
I have problems starting with a closeup portrait but want a full body picture as an end result. Playground AI sticks on the upper body. Can you give me a hint, please?
That's a bit tricky since there is no outpainting feature but there is a work around. You can download the image, recompose it in another editor like photoshop, pixlr or even Canva. Then re-upload and use inpaint to add the rest. I'll be doing a video on this soon. It's not a fool proof method but it can work to some extent.
I see that 1.7K views. 💯⚡️
Thanks bud, I knew the community would like the topic but I didn't realize how much! ☺🙌
How did you get those peresets? I don't see any of them in my menu.
Hey bud, any chances you are using the Playground V1 model or Stable Diffusion 2.1? All the filters are available with Stable Diffusion 1.5 at the moment. Only select ones for Playground V1 model. Let me know if that helps!
HELLO, FIRST, EXCELENTS TUTORIALE, BUT, WHY I CANT USE IN PAINT THANKS gracias amigos
You’re welcome! Most likely you have to turn the filter off before you can inpaint. 👍🏼
can this creating your own avatar be done in leonardo ai? maybe you should do a video on that as well!
Hello again Julia! Yes absolutely and I am preparing a video for that as well. However the benefit of Leonardo AI is that you can train your own model based on real people if you have the photos to use for training. Check it out here th-cam.com/video/oNajDvx3_GY/w-d-xo.html and I also cover some updates on this follow up video on the second half of the video th-cam.com/video/ysg06oLGA94/w-d-xo.html
i dont understand why i cant select impaint
Most likely you have the filter on, turn that off and you will be able to use inpainting. 👍
@@MonzonMedia thank you, that was the problem.
@@beatprocess you’re welcome! Sorry for not making it clear in the video. Didn’t really realize that until now. So thank you! 👍🏼
When I open my FILTER, the samples are different from yours. My question is: How can I upload other samples of images to change the styles in my filter?
Make sure you are using stable diffusion 1.5 as your model, then you'll be able to use all the filters that are available. Hope that helps!
@@MonzonMedia Thank you, very appreciated.
How do you make a cartoon/anime character look realistic?
I’m assuming you are using a reference image? You can start with a filter like analog style, in the negative prompt area you can put in cartoon, 2D, drawing etc. In your main prompt it can help to put photo of, portrait of, photorealistic. Might take a few tries. Hope that helps!
Is there a difference between using a mask or inPaint?
Great question! Technically, they do the same thing, isolate an area. I tend to say mask more because that's term we use for photo editing. But in the AI world we say inpaint. I'm sure others would have their own opinion on it but at the end of the day, it's the same thing. Isolating an area. Hope that makes sense?
@@MonzonMedia thank you for your quick answer
@@Grimmona My pleasure, glad to help!
didnt work, I followed the entire tutorial, but the free version is limit! my result everything is mesed up.
Sometimes it takes the right image. I have yet to find an image that can do me well...just looks weird hahaha! You can keep trying with another image and works best with head and shoulder or half body shots. It's much better to be able to train a model with many images so that you can get the likeness of the person easier. Have you tried Leonardo AI? They have the ability to train. I did a video on it here. th-cam.com/video/oNajDvx3_GY/w-d-xo.html and at the end of this video I gave some updates. th-cam.com/video/ysg06oLGA94/w-d-xo.html. The free tier is a bit limited though.
Hello and thanks for a comprehensible video. Actually I am going crazy with this, as it is AI it should have a logical pattern of behaving. I did created a picture of a blueberry monster B for Berry and it was supposed to hold in the hand a banana. He was not, the bananas were on the ground and one above the head. I could erase them easily but then I spent five hours trying to add it to the picture so it would be in his hand. NO way. I was able to put bananas in different places, to place different objects in the picture. I tried image to image, inpainting and I am kinda deceived because I could do it easily in ten minutes in Photoshop and I lost all my day, in expectation that I would learn how to do adding and removing easily. Do you have some video on this topic, like detailed one, dealing with what can be established wrongly if it is so difficult to add a simple banana to the already perfect picture.
Thanks for checking out the video and the feedback. AI in general has trouble holding objects in hands. Usually inpainting can fix it or get close to what you want it to look like but I've found that it's sometimes quicker to do it in post. Also, many of the web based platforms lack certain features like ControlNet for example which you can run locally on a PC and this allows you make specific poses. I'd love to do a video on this topic actually.
@@MonzonMedia So, I will be impatiently waiting for the video, many thanks in advance
Super🙏🙏 sir
Hello my friend! Always a pleasure to see ya! 🙏👍
I prefer the real April.
😁
(Great tutorial though)
Hahaha! I do too! I have her booked for a spring shoot. You won’t catch me out there during the winter hahaha! 👍🏼
Unless I upgrade my plan, or else I can only do simple editing.
There isn't much of a difference in how the product performs between the free and paid options, mostly it's the step count (quality) that is limited to 50 which is more than enough for most images.
is there a way to do this in leonardo ai?
absolutely! going to cover this soon! Kind of the same process really.
"Inpaint" is unactive for clicking. What can be wrong?
UPD: Ah, got it. Just need to deactivate the filter.
Yup you got it!
Then instructions seemed pretty straight forward, but my first image was NSFW even thought it was just a regular pic of me in a gray t-shirt. Then the generate button stopped working. Tried to reload but its like it doesn't exist. Oh well.
Perhaps try another image?
this didnt help, it wasnt working for me, the program has changes a little since this video
Are you creating on board or canvas? There hasn’t been too many changes on board other than new features on canvas
I tried but why the inpaint button cannot be clicked?
Most likely the filter is on, you have to turn it off before inpainting. 👍
why is that i can't generate image? the button is just in grey color?
Hi JP, some context would help.
if I try this with my own picture (this one), I just get an ugly copy paste on something generated.
try another picture, it also helps to bring the image strength up a little at a time.
Can i use it on Android smartphone?
As far as I know yes, however it may not be optimized for mobile.
@@MonzonMedia the webpage doesn't even responds on phone. I don't know what to do. I want to upload my photo and make self portrait
Healing an image, like remove some spot from an image a bit heard and problematic.
Have you tried the object eraser in canvas? Works pretty well.
After generating a good image, how to insert/replace my own face in that image?
I go through it in the video. You need to start with your image first.
inpaint isnt available any more
Sure it is, I use it all the time. After you generate an image on the top corner is your actions menu, there you will see inpaint. If it's greyed out you have to turn off the filter.
@@MonzonMedia thank you will try it. mine was grayed out.
😍❤💯👍
Glad you find it helpful! 🙌👍
Can you please give me negative prompt of things?
Should be a link in the description my friend. 👍🏼
@@MonzonMedia I saw it,it just has a description for people
I asked if there is a negative prompt for cars,toys,things etc...
aaahhhh I see, well in that case just use words that you don't want to see in the image. Whatever results are not to your liking just add it as you go, like blurry, warped (for things like tires). It's a bit difficult to come up with one list for objects as there are many different objects.
@@MonzonMedia ok bro,thanks a lot ❤
Prompt: "cartoon"
vector, 2d illustration, comic...lot's of choices 😁
here is the negative things?
Do you mean the negative prompts? It's in the description.
thanks
@@geniune168 No problem! 👍
She's still blind...
🤔Sorry not sure what you mean?
playground is garbage lol
What makes you come to that conclusion? I like it personally but it may not be for everyone.